Web Analytics Segment Contribution Scoring

The patent applies segmentation by dividing the web analytics metric into multiple segments based on visitor attributes (e.g., new vs. returning visitors, device types, geographic locations). Each segment is independently analyzed to determine its contribution to metric changes, making the complex identification process manageable through structured division of the data space.

The patent introduces segment contribution scores as an intermediary metric that mediates between raw visitor data and the final identification of contributing segments. These scores quantify the impact of each segment on metric changes, serving as a computational bridge that simplifies the analysis process while preserving complete information from all segments.

The patent changes the parameter of analysis by computing segment contribution scores that directly measure the impact of each segment on metric changes. This parameter transformation allows the system to accurately identify contributing segments without exhaustively analyzing all segments in detail, as the scores prioritize segments by their actual contribution to the observed changes.

The patent applies partial action by focusing computational resources on segments that have sufficient visitor data and are likely to contribute significantly to metric changes. Rather than uniformly analyzing all segments with equal depth, the system selectively deepens analysis for high-impact segments while using aggregated data for others, reducing overall analysis time while maintaining accuracy.

The patent transforms the computational problem by changing from detailed analysis of all visitor attributes to computing aggregated segment contribution scores. This parameter change reduces computational complexity by working with pre-grouped segments rather than individual visitor records, achieving precise segment identification with significantly reduced computational resources.

The patent applies partial action by computing segment contribution scores only for segments that meet certain criteria (e.g., sufficient visitor data volume, significant expected impact). This selective computation approach maintains precision for the most relevant segments while avoiding unnecessary computational expenditure on segments unlikely to contribute meaningfully to metric changes.